Tutorial for 2d cores added

[neuronmorph]

Tutorial provides examples on how to use different 2d cores for sensorium 2022

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 7.51%. Comparing base (1323caa) to head (185cd90).

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #233   +/-   ##
=====================================
  Coverage   7.51%   7.51%           
=====================================
  Files         58      58           
  Lines       5978    5978           
  Branches    1018    1018           
=====================================
  Hits         449     449           
  Misses      5492    5492           
  Partials      37      37           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pollytur]

Based on discussion with Konsti

• small configs for each core type, highlighting the differences
• core_config not model_config
• change a bit the description for the SE core
• should loading and not loading a ppretrain core for transfer learning

[KonstantinWilleke]

Very nice PR! I have another request that would be nice to add:
Could you add to the top of the notebook a small description of Sensorium and the dependencies.

So that you link to the sensorium 2022 repo so that it is clear which dataset is being used. And there is also the dependecy on the sensorium package:
dataset_fn = 'sensorium.datasets.static_loaders'
This line requires that the sensorium package is installed.

A possible solution is: you can describe that if they have sensorium installed, they can use this data. But in case the user wants to proceed without it, you can just initialize a torch.tensor (like torch.randn(32,1,144,256)) which has the same size. And then the user does not need "real" images, but just a tensor that is essentially the same.

[pollytur]

@KonstantinWilleke
The installation is there

But I think the import (like import sensorium is missing)

I actually would disagree with the explanation of sensorium dependencies, as it's not the goal of this tutorial- the goal is to show the cores. And I expect for some people to be very distracted by this from the beginning.

Using random inputs is actually a great idea and would make things easier and more disentangled. This way we do not need either installation or explanations.

[pollytur]

Could you please update it with the random input for the cores as Konsti suggested (removing the dataset creation and sensorium installation) and then we could merge this great example?

[neuronmorph]

Please have a look. I removed the Sensorium 2022 dependencies.

[pollytur]

Great, thanks a lot!

[KonstantinWilleke]

Great work @neuronmorph! Thanks a lot. You can merge the PR - all that is left to do is to update the branch. You can click the button "update branch" and then feel free to merge.